Hydraulic excavator track chassis dynamic performance prediction method based on virtual test field

Abstract

This invention relates to the field of engineering machinery simulation technology, and more particularly to a method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field. First, a high-precision virtual prototype model of the tracked chassis and a discrete element ground model are established. Second, using the measured angle of repose as the response target, significant contact parameters are determined, and the optimal combination of significant contact parameters is obtained by using a GA-BP neural network. Finally, a discrete element ground in the virtual test environment is constructed using the optimal significant contact parameters, and the virtual prototype model is subjected to motion experiments on the discrete element ground, with real-time measurement of the dynamic performance of the virtual prototype model. This invention significantly improves the confidence level of virtual simulation and provides an efficient and reliable analysis platform for predicting and optimizing the dynamic performance of chassis systems under complex extreme working conditions.

Description

Technical Field

[0001] This invention belongs to the field of engineering machinery simulation technology, and in particular relates to a method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field. Background Technology

[0002] Hydraulic excavators are key equipment in modern infrastructure, mining, and emergency rescue. The dynamic performance of their tracked chassis system directly determines the machine's work efficiency, reliability, and service life. In actual operation, the tracked chassis frequently encounters complex conditions such as rugged roads, slopes, and turns. The unsteady interaction between the chassis and the ground generates strong impact and vibration loads. These loads can easily lead to premature failure of the track roller bearings and structural damage such as crack propagation in the track shoes, seriously affecting the equipment's service safety and lifespan.

[0003] Currently, research on the dynamic performance of tracked chassis mainly relies on physical testing methods. By installing sensors (such as accelerometers and torque meters) on physical prototypes, dynamic load spectra under certain operating conditions can be obtained. However, physical testing methods have limitations such as high testing costs and long testing cycles, difficulty in reproducing extreme and complex operating conditions, and incomplete data acquisition.

[0004] To overcome the bottlenecks in physical testing, virtual test field technology has emerged. This technology integrates multi-physics coupling simulation methods such as multibody dynamics, discrete element method, and finite element method to construct high-fidelity system models in computers, thereby achieving low-cost, high-efficiency, and repeatable simulation analysis of various working conditions.

[0005] Currently, the application of virtual test track technology in the field of wheeled vehicles is relatively mature and has been successfully used for performance prediction such as vehicle durability, fatigue analysis, and collision safety. However, applying virtual test track technology to tracked construction machinery, especially hydraulic excavators, still faces the problem of the lack of a high-precision track-ground interaction model: the contact between tracked vehicles and the ground is a continuous, complex dynamic process with large deformations. Its simulation accuracy heavily depends on the accuracy of contact parameters (such as static friction coefficient, rolling friction coefficient, etc.) in the discrete element ground model. These parameters are difficult to obtain through theoretical calculations, and traditional parameter calibration methods (such as trial and error methods and simple response surface methods) suffer from low efficiency and are prone to getting trapped in local optima, resulting in the constructed discrete element ground model failing to truly reflect the mechanical properties of the actual road surface.

[0006] Therefore, there is an urgent need in this field for a method that can accurately calibrate discrete element ground contact parameters and build a high-confidence virtual test field on this basis, so as to accurately predict the dynamic response of hydraulic excavator tracked chassis under complex working conditions, thereby guiding design optimization and reducing R&D costs and risks. Summary of the Invention

[0007] In view of this, the present invention aims to provide a method for predicting the dynamic performance of hydraulic excavator track chassis based on a virtual test field. By calibrating discrete element ground contact parameters with high precision, and constructing a high-confidence virtual test field on this basis, the authenticity of track-ground interaction in the virtual environment is ensured.

[0008] To achieve the above objectives, the technical solution created by this invention is implemented as follows:

[0009] A method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field includes:

[0010] S1: Establish a virtual prototype model of the excavator tracked chassis; determine the ground contact parameters in the test environment where the excavator is located, and measure the actual angle of repose of the ground;

[0011] S2: Select significant contact parameters that have a significant impact on the angle of repose of the ground from the contact parameters obtained in step S1; use the actual angle of repose as the search optimization target and use the GA-BP neural network to determine the optimal significant contact parameters.

[0012] S3: Select the optimal saliency contact parameters determined in step S2 to construct the discrete element ground in the virtual test environment, and then conduct motion experiments on the virtual prototype model established in step S1 on the discrete element ground to measure the dynamic performance of the virtual prototype model in real time.

[0013] Furthermore, the process of establishing a virtual prototype model of the excavator tracked chassis in step S1 includes: performing simulation modeling on each component of the excavator tracked chassis, and determining the constraints and contact relationships between each simulation modeling component to obtain a virtual prototype model.

[0014] Furthermore, in step S1, the ground material is granite. The process of measuring the actual angle of repose of the ground includes: loading granite blocks into a bottomless cylinder with an alloy steel plate at the bottom; vertically lifting the cylinder to allow the granite blocks to naturally accumulate on the alloy steel plate, and photographing the accumulated granite blocks; extracting the outermost accumulation contour of the granite blocks from the photographed accumulation image, obtaining the contour curve coordinates, and using nonlinear fitting to obtain a Gaussian distribution function; using the highest point and curve width in the Gaussian fitting curve, calculating the actual angle of repose using the following formula:

[0015] ;

[0016] Where θ represents the actual angle of repose, y max σ represents the ordinate of the highest point, and σ represents the width of the curve.

[0017] Furthermore, in step S2, the actual angle of repose is used as the search optimization objective, and the optimal significant contact parameter is determined using a GA-BP neural network. This process includes: establishing a BP neural network with the significant contact parameter as input and the angle of repose as output; using the BP neural network as the fitness function of the genetic algorithm, using the measured angle of repose as the search optimization objective of the genetic algorithm, back-calculating the significant contact parameter, determining the optimal solution of the significant contact parameter, and obtaining the optimal significant contact parameter.

[0018] Furthermore, step S2 also includes: determining the optimal combination of significant contact parameters, and fitting a regression equation between the significant contact parameters and the angle of repose using the optimal combination; substituting the actual angle of repose into the regression equation to obtain the significant contact parameter corresponding to the fitted angle of repose that is closest to the actual angle of repose; calculating the relative error between the fitted angle of repose and the actual angle of repose, and calculating the relative error between the angle of repose corresponding to the optimal significant contact parameter and the actual angle of repose; comparing the two relative errors, and selecting the significant contact parameter corresponding to the angle of repose with the lower relative error for constructing the discrete element ground in step S3.

[0019] Compared with the prior art, the present invention can achieve the following beneficial effects:

[0020] This invention presents a method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field. It is the first to systematically apply GA-BP neural network parameter inversion technology to the construction of a virtual test field for a hydraulic excavator tracked chassis. By introducing a GA-BP neural network and using the measured angle of repose as the response target for parameter inversion, significant contact parameters are selected. This effectively solves the problem of traditional response surface methodology easily getting trapped in local optima in nonlinear problems, significantly improving the confidence level of the virtual test field. It provides an efficient and reliable analysis platform for predicting and optimizing the dynamic performance of chassis systems under complex extreme conditions. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0022] Figure 1 A flowchart illustrating the dynamic performance prediction method for hydraulic excavator tracked chassis based on a virtual test field, as described in an embodiment of the present invention;

[0023] Figure 2 A flowchart illustrating the dynamic performance prediction method for hydraulic excavator tracked chassis based on a virtual test field, as described in an embodiment of the present invention.

[0024] Figure 3The graphs showing the influence of the static friction coefficient between granite and granite, the static friction coefficient between granite and structural steel, and the rolling friction coefficient between granite and granite on the angle of repose, as described in the embodiments of the present invention;

[0025] Figure 4 A schematic diagram of the BP neural network described in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof.

[0027] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0028] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0029] The invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0030] like Figure 1 As shown in the embodiment of the present invention, the method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field includes:

[0031] S1: Establish a virtual prototype model of the excavator tracked chassis; determine the ground contact parameters in the test environment where the excavator is located, and measure the actual angle of repose of the ground. In some embodiments, the ground material is granite. In this embodiment of the invention, the contact parameters include the contact parameters between the granite and the track plates, and the contact parameters between granite pieces, including but not limited to the coefficient of restitution, the coefficient of static friction, and the coefficient of sliding friction.

[0032] In some embodiments, the process of establishing a virtual prototype model of the excavator tracked chassis in step S1 includes: performing simulation modeling on each component of the excavator tracked chassis, and determining the constraints and contact relationships between each simulation modeling component to obtain a virtual prototype model.

[0033] In this embodiment of the invention, a virtual prototype model of a hydraulic excavator tracked chassis is established based on the multibody dynamics software Recurdyn, specifically including:

[0034] The 3D model of the hydraulic excavator was simplified in SolidWorks, exported to STEP format, and then imported into the dynamics simulation software. The simulation focused on the tracked chassis; therefore, components such as the superstructure and cab were slightly simplified to ensure their center of gravity was consistent with reality.

[0035] Based on the actual dimensions of the drive wheel, guide wheel, track roller, carrier roller and track link of the hydraulic excavator track chassis, the above components are modeled in the low-speed track module of Recurdyn, and the material properties and mass are defined. There are 9 track rollers on each side of the hydraulic excavator.

[0036] Constraints are defined between each wheel and axle, and between the axle and track frame. Based on the track chassis tensioning device, the spring stiffness and damping are defined to simulate the track tensioning effect of the track tensioning device on the track.

[0037] In some embodiments, the actual angle of repose of the ground is measured using the lifting cylinder method, the process of which includes:

[0038] Granite blocks are placed into a bottomless cylindrical container, with an alloy steel plate at the bottom of the container.

[0039] A vertically lifted cylinder allows granite blocks to accumulate naturally on an alloy steel plate, and the accumulated granite blocks are then photographed.

[0040] Extract the outermost stacked contour of the granite blocks from the captured stacked images, obtain the contour curve coordinates, and use nonlinear fitting to obtain the Gaussian distribution function;

[0041] Using the highest point and the width of the Gaussian fitted curve, the actual angle of repose can be calculated using the following formula:

[0042] ;

[0043] Where θ represents the actual angle of repose, y max σ represents the ordinate of the highest point, and σ represents the width of the curve.

[0044] In this embodiment of the invention, the process of measuring the actual angle of repose of the ground specifically includes: loading 1 kg of granite blocks into a bottomless cylinder with an inner diameter of 200 mm and a height of 600 mm, with the bottom of the cylinder placed on a 500 mm × 500 mm alloy steel plate. The cylinder is vertically lifted at a speed of 0.3 m / s, allowing the blocks to accumulate naturally, and the accumulated granite blocks are photographed. Grayscale processing and binarization methods are applied to the accumulated image to reduce image noise. The Sobel operator is used to perform edge detection on the noise-reduced accumulated image to extract the outermost accumulated contour of the granite blocks. Finally, Origin software is used to obtain the image contour curve coordinates, and a Gaussian distribution function is obtained using nonlinear fitting. The granite's angle of repose is calculated using the highest point and curve width of the Gaussian fitted curve, according to the above formula. This process is repeated ten times, each time taking accumulated images from both the front and side views to obtain the average angle of repose of the granite. This average angle of repose is the actual angle of repose. In this embodiment of the invention, the actual angle of repose is 28.55°, with an error of ±1.33°.

[0045] Furthermore, in this embodiment of the invention, the material parameters of the granite are measured, specifically including:

[0046] Granite density: Granite was processed into cylindrical specimens with a diameter of 50 mm and a height of 100 mm. The specimens were weighed using an electronic scale with an accuracy of 0.01 g, and the average density was calculated to be 2.56 g / cm³.

[0047] Poisson's ratio and Young's modulus: A uniaxial compression test was conducted on a cylindrical specimen using a universal testing machine at a loading rate of 2 mm / min. Based on the axial and radial strain data, the Poisson's ratio was found to be 0.263 and the Young's modulus to be 61.5 GPa.

[0048] Recovery coefficient range: Using a free fall test, a granite block was dropped from a height of 200mm onto an alloy steel plate (the same material as the track plate), and the rebound height was measured. The results of 20 tests showed a range of 0.27-0.48.

[0049] S2: Select significant contact parameters that have a significant impact on the angle of repose of the ground from the contact parameters obtained in step S1; use the actual angle of repose as the search optimization target and use the GA-BP neural network to determine the optimal significant contact parameters.

[0050] In this embodiment of the invention, the contact parameters of granite were screened for significance factors using the Plackett-Burman experiment. Six contact parameters with potential influence were selected as shown in Table 1. Their level ranges were determined based on previous experiments and literature. The contact parameters are shown in Table 1.

[0051] Table 1: Six contact parameters:

[0052]

[0053] In Table 1, +1 represents the parameter value at the highest contact level, and -1 represents the parameter value at the lowest contact level. The granite-granite restitution coefficient (A in Tables 1-3) is specifically determined by a free-fall test, where a granite block is dropped freely from a height of 200 mm onto another granite block, and the rebound height of the granite block is measured. The granite-granite restitution coefficient ranges from 0.25 to 0.55.

[0054] The granite-structural steel recovery coefficient (B in Tables 1-3) is specifically determined by using a free-fall test, where a granite block is dropped freely from a height of 200mm onto a structural steel plate (material similar to the track plate), and the rebound height of the granite block is measured to obtain the granite-structural steel recovery coefficient, which ranges from 0.27 to 0.48.

[0055] The static friction coefficient between granite blocks (C in Tables 1 to 3) is determined by placing a granite block on a granite slab and performing an inclined plane sliding test. The inclination angle of the granite slab is slowly increased until the granite block begins to slide. The tangent of this angle is the static friction coefficient between granite blocks, ranging from 0.3 to 0.75.

[0056] The static friction coefficient between granite and structural steel (D in Tables 1-3) is determined by placing a granite block on a structural steel plate (made of the same material as the track plate) and conducting an inclined plane sliding test. The inclination angle of the structural steel plate is slowly increased until the granite block begins to slide. The tangent value of this angle is the static friction coefficient between granite and structural steel, ranging from 0.42 to 0.7. The rolling friction coefficient between granite and granite (E in Tables 1-3) is determined by rolling a granite block at a constant speed on an inclined granite slab. At this point, the gravitational component of the granite block is balanced with the rolling resistance. The tangent value of this angle is the rolling friction coefficient between granite and granite, ranging from 0.05 to 0.25.

[0057] The rolling friction coefficient of granite-structural steel (F in Tables 1 to 3) is specifically determined when a granite block rolls at a constant speed on an inclined structural steel plate. At this point, the gravitational component of the granite block is balanced with the rolling resistance. The tangent of the angle at this point is the rolling friction coefficient of granite-structural steel, which ranges from 0 to 0.2.

[0058] An orthogonal experimental design was generated using Design-Expert, and a lifting cylinder simulation experiment was conducted in the EDEM discrete element model software. The angle of repose was measured, and the results were analyzed using Plackett-Burman orthogonal experimental design. A p-value less than 0.01 in the variance analysis indicates that the influencing factors have a highly significant impact on the objective function; a p-value less than 0.05 indicates that the influencing factors have a relatively significant impact on the objective. As shown in Table 2, the variance analysis results show that the granite-granite rolling friction coefficient E has a highly significant impact on the angle of repose, while the granite-granite static friction coefficient C and the granite-structural steel static friction coefficient D have significant impacts. All three contact parameters have positive feedback effects. Figure 3 As shown, where Figure 3 Figures (a), (b), and (c) show that the static friction coefficient C between granite and granite, the static friction coefficient D between granite and structural steel, and the rolling friction coefficient E between granite and granite have a significant impact on the angle of repose. Figure 3 The solid black line represents the angle of repose, and the area enclosed by the dashed black line represents the 95% confidence interval. As the static friction coefficients of granite-granite, rolling friction coefficient, and granite-structural steel static friction coefficient increase, the angle of repose increases. The static friction coefficients of granite-granite and granite-structural steel have relatively small effects on the angle of repose, changing slowly within the optimal range. The rolling friction coefficient of granite-granite has the most significant effect on the angle of repose. Therefore, the significant contact parameters include the granite-granite static friction coefficient C, the granite-structural steel static friction coefficient D, and the granite-granite rolling friction coefficient E. The other three contact parameters have no significant effect on the angle of repose and are therefore not considered significant contact parameters.

[0059] Table 2: Results of Variance Analysis of Plackett-Burman Orthogonal Experiment

[0060]

[0061] In Table 2, "Model" represents the statistical variance analysis model obtained from the Plackett-Burman orthogonal experiment, and "d" represents the model obtained from the statistical variance analysis. f"" represents the number of variables that can be freely varied in the statistical model of the analysis of variance, "F" represents the strength of the influence of the independent variable parameter, and "P" represents whether the factor is statistically significant; the smaller the value, the more statistically significant it is. Based on the results of the Plackett-Burman orthogonal experimental analysis of variance, three significant factors (granite-granite static friction coefficient C, granite-structural steel static friction coefficient D, and granite-granite rolling friction coefficient E) were used for the steepest slope test design, and the remaining insignificant factors were taken at the median level. The optimal combination of significant contact parameters was determined through the Central Composite experiment, as shown in Table 3. The analysis results show that the interaction between the granite-granite static friction coefficient and the rolling friction coefficient, and between the granite-granite static friction coefficient and the granite-structural steel static friction coefficient, and the square term of the three factors have a highly significant effect on the angle of repose, and the granite-structural steel static friction coefficient has a significant effect on the angle of repose.

[0062] Table 3: Experimental Analysis Results of Central Composite

[0063]

[0064] In some embodiments, step S2 uses the actual angle of repose as the search optimization objective and utilizes a GA-BP neural network to determine the optimal significant material parameters. This process includes:

[0065] A BP neural network is established with significant material parameters as input and the angle of repose as output;

[0066] Using a BP neural network as the fitness function of a genetic algorithm and the measured angle of repose as the search and optimization objective of the genetic algorithm, the salient material parameters are back-calculated to determine the optimal solution of the salient material parameters, thus obtaining the optimal salient material parameters.

[0067] In this embodiment of the invention, the BP neural network is as follows: Figure 4 As shown, the network consists of an input layer with 3 neurons, a hidden layer with 8 neurons, and an output layer with 1 neuron. The 3 neurons in the input layer correspond to the static friction coefficients E of granite-granite, granite-structural steel, and granite-granite rolling friction coefficients, respectively. The neuron in the output layer corresponds to the predicted angle of repose. The BP network training algorithm is the Levenberg-Marquardt backpropagation algorithm, with the training performance function being the mean squared error. The learning rate is 0.001, and the maximum number of training steps is set to 50. Each significant material parameter is taken from the -1, 0, 1 levels of Central composite and arranged in various combinations. Three sets of zero-level estimation errors are added, resulting in a total of 30 sets as the neural network dataset. The training, validation, and test sets are divided in a 7:1.5:1.5 ratio to train the BP neural network.

[0068] This invention employs a genetic algorithm to iteratively optimize an individual population by simulating selection, crossover, and mutation in a biological genetic system, thereby improving the algorithm's local search capability and preventing it from getting trapped in local optima. In this embodiment, the genetic algorithm undergoes 300 evolutionary iterations, the population size is 100, the selection function is normalGeomSelect with a coefficient of 0.09, the crossover coefficient is 0.8, and the mutation coefficient is 0.2.

[0069] This invention utilizes the nonlinear optimization of a genetic algorithm, using a BP neural network model as the fitness function and the measured angle of repose as the search optimization objective. It then back-calculates the optimal significant material parameters based on the input parameters: the static friction coefficient (C) between granite and granite, the static friction coefficient (D) between granite and structural steel, and the rolling friction coefficient (E) between granite and granite. The corresponding mean angle of repose is 28.74°, with a relative error of 0.67%.

[0070] In some embodiments, step S2 further includes:

[0071] The optimal combination of significant contact parameters was determined by the Central composite test, and the regression equation between significant contact parameters and the angle of repose was constructed by fitting the optimal combination.

[0072] Calculate the relative error between the fitted angle of repose and the actual angle of repose, and calculate the relative error between the angle of repose corresponding to the optimal saliency contact parameter determined by the BP neural network and the actual angle of repose;

[0073] The two relative errors are compared, and the significant contact parameter corresponding to the repose angle with the lower relative error is selected for constructing the discrete element ground in step S3.

[0074] In this embodiment of the invention, the RSM response surface methodology is used to fit and construct three significant contact parameters and the angle of repose θ using the optimal combination obtained from the Central Composite experimental analysis results. AOR The regression equation between them is as follows:

[0075] θ AOR =30.30+1.28C+0.6633D+3.34E+0.8887CE

[0076] -1.14C 2 -0.7851D 2 -0.8276E 2 ;

[0077] Substituting the actual angle of repose multiple times into the regression equation yields several corresponding fitted angles of repose. The average of these fitted angles of repose is the final fitted angle of repose, specifically 29.04°, with a relative error of 1.73% compared to the actual angle of repose. The results show that the simulated angle of repose obtained using the GA-BP method has a smaller error than that obtained using the RSM method and is closer to the measured angle of repose of granite. The optimal significant contact parameters determined by GA-BP are used as the design attribute parameters for the subsequent virtual test field ground.

[0078] S3: Select the optimal saliency contact parameters determined in step S2 to construct a discrete element ground in the virtual test environment. Then, conduct motion experiments on the virtual prototype model established in step S1 on the discrete element ground and measure the dynamic performance of the virtual prototype model in real time. In this embodiment of the invention, using the EDEM discrete element model software, a discrete element ground with a length of 18m and a width of 6m is constructed by selecting the optimal saliency contact parameters determined in step S2. Through the coupling interface between Recurdyn and EDEM, the virtual prototype model and the discrete element ground are coupled to simulate straight-line driving, climbing, and turning.

[0079] To verify that the method provided by this invention can realistically simulate the movement of an excavator on a granite surface and ensure the authenticity of the track-ground interaction in the virtual environment, this invention also conducted a real-machine test. In the real-machine test, a real hydraulic excavator was used to perform straight-line, turning, and climbing multi-condition walking tests. Encoders, an RTK positioning system, displacement sensors, and tilt sensors were used to measure the track speed, tensioning device displacement, and overall excavator tilt angle during the hydraulic excavator's movement, setting the driving conditions and initial state information for the virtual prototype model.

[0080] Furthermore, the drive torque of the hydraulic excavator's drive wheels and the frequency domain response of the track roller acceleration were measured to verify the accuracy and reliability of the virtual test field model. The drive torque is closely related to factors such as the track speed, tension, and road slope, and is also significantly affected by the parameter settings between the track roller and the track plate, as well as the parameter settings between the track plate and the ground. It is a key comparative factor in ensuring the accuracy of the dynamic model.

[0081] The driving force of a hydraulic excavator mainly comes from the pressure and flow provided by the hydraulic system. A pressure sensor is connected to the forward main pressure test port of the hydraulic motor to monitor pressure changes in the hydraulic system in real time. The pressure changes at the oil measuring port are measured in real time under different operating conditions, and the measurement data is transmitted to the data acquisition system. Based on the conversion relationship between hydraulic motor flow rate, pressure, and torque, combined with the hydraulic system flow rate (Q) and hydraulic motor efficiency (η), and given that the transmission ratio of the motor reducer is 76.44, the driving torque of the drive wheel is finally calculated using the following formula:

[0082] ;

[0083] Where T represents the driving torque and n represents the transmission ratio of the motor reducer.

[0084] The measured and simulated driving torques were compared using mean error and extreme error analysis. Mean error was analyzed using the mean of the driving torque data, while extreme error was analyzed using envelope error analysis. Envelope error analysis involves analyzing the outer contour curve of the data. In the time domain, the envelope refers to the local maximum (upper envelope) or minimum (lower envelope) of the signal. The original signal was transformed into a complex signal using Hilbert transform, where the real part is the original signal and the imaginary part is the "sine" component. The envelope was obtained by calculating the amplitude of the complex signal, and the instantaneous amplitude and phase of the signal were extracted, as shown below:

[0085] ;

[0086] Where H(μ(t)) represents the signal after performing the Hilbert transform on the original signal μ(t), P represents the pressure detected by the real-time monitoring hydraulic system, and τ represents the integral element.

[0087] Using the Hilbert-transformed signal H(μ(t)) as the imaginary part and the original signal μ(t) as the real part, we obtain the complex signal Z(t) = μ(t) + iH(μ(t)), where i represents the imaginary unit.

[0088] The amplitude of a complex signal gives the envelope of the signal, and the phase of a complex signal gives the instantaneous phase of the signal. Thus, the instantaneous amplitude and instantaneous phase are obtained, as follows:

[0089] ;

[0090] ;

[0091] Where A(t) and θ(t) represent instantaneous amplitude and instantaneous phase, respectively.

[0092] As a key support component of the track system, the track roller's acceleration reflects the vibration and impact it experiences during operation. By measuring the frequency domain response of acceleration, the dynamic response characteristics of the track roller can be evaluated. Accelerometers measure the acceleration of the track roller in various directions, and the data is transmitted to a computer via a data acquisition unit to obtain vibration signals. Fourier transform is used to convert the acceleration time-domain signal into a frequency-domain analysis, thereby performing acceleration power spectrum analysis. The formula for Fast Fourier Transform (FFT) is:

[0093] ;

[0094] Where X(k) represents the frequency domain signal, x(m) represents the time domain signal, and M represents the total time.

[0095] After comparing the original drive torque data, Hilbert transform was used for local extremum error analysis. Envelope curves were extracted from the measured drive torque in the actual machine test and the drive torque in the virtual prototype model's driving simulation experiment on a discrete-element ground using the method provided in this invention. The relative error of the mean values ​​of the two drive torques was less than 10%, and the relative error of the two envelope analyses was less than 15%. During the actual machine test and the process provided by this invention, the accelerometer data of the track's rearmost support roller was subjected to a Fast Fourier Transform to obtain the acceleration power spectrum. Comparing the two acceleration power spectra, it was found that the frequency domain characteristics of the support roller acceleration were highly similar, with a main frequency amplitude deviation of less than 2dB and a frequency band coverage of greater than 90%. This demonstrates that the method provided by this invention can realistically simulate the movement of an excavator on a granite surface, effectively ensuring the realism of the track-ground interaction in the virtual environment.

[0096] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0097] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field, characterized in that, include: S1: Build a virtual prototype model of the excavator tracked chassis; Determine the contact parameters between the excavator and the ground in the test environment, and measure the actual angle of repose of the ground; S2: Select significant contact parameters that have a significant impact on the angle of repose of the ground from the contact parameters obtained in step S1. Using the actual angle of repose as the search optimization objective, the optimal salience contact parameters are determined using a GA-BP neural network; In step S2, the actual angle of repose is used as the search optimization objective, and the optimal salient contact parameter is determined using a GA-BP neural network. This process includes: establishing a BP neural network with the salient contact parameter as input and the angle of repose as output. Using a BP neural network as the fitness function of a genetic algorithm and the measured angle of repose as the search and optimization objective of the genetic algorithm, the salient contact parameters are back-calculated to determine the optimal solution of the salient contact parameters, thus obtaining the optimal salient contact parameters. S3: Select the optimal saliency contact parameters determined in step S2 to construct the discrete element ground in the virtual test environment, and then conduct motion experiments on the virtual prototype model established in step S1 on the discrete element ground to measure the dynamic performance of the virtual prototype model in real time.

2. The method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field according to claim 1, characterized in that, The process of establishing a virtual prototype model of the excavator tracked chassis in step S1 includes: performing simulation modeling on each component of the excavator tracked chassis, determining the constraints and contact relationships between each simulation modeling component, and obtaining the virtual prototype model.

3. The method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field according to claim 1, characterized in that, In step S1, the ground material is granite, and the process of measuring the actual angle of repose of the ground includes: Granite blocks are placed into a bottomless cylindrical container, with an alloy steel plate at the bottom of the container. A vertically lifted cylinder allows granite blocks to accumulate naturally on an alloy steel plate, and the accumulated granite blocks are then photographed. Extract the outermost stacked contour of the granite blocks from the captured stacked images, obtain the contour curve coordinates, and use nonlinear fitting to obtain the Gaussian distribution function; Using the highest point and the width of the Gaussian fitted curve, the actual angle of repose can be calculated using the following formula: ； Where θ represents the actual angle of repose, y max σ represents the ordinate of the highest point, and σ represents the width of the curve.

4. The method for predicting the dynamic performance of a hydraulic excavator tracked chassis based on a virtual test field according to claim 1, characterized in that, Step S2 also includes: Determine the optimal combination of significant contact parameters, and fit the optimal combination to construct a regression equation between significant contact parameters and the angle of repose; Substituting the actual angle of repose into the regression equation, we obtain the significant contact parameter corresponding to the fitted angle of repose that is closest to the actual angle of repose. Calculate the relative error between the fitted angle of repose and the actual angle of repose, and calculate the relative error between the angle of repose corresponding to the optimal significant contact parameter and the actual angle of repose; The two relative errors are compared, and the significant contact parameter corresponding to the repose angle with the lower relative error is selected for constructing the discrete element ground in step S3.

Images